>Rich Measures wrote:
>>If the electronic protection circuit shorts during a major fault, which
>>hardly seems to be uncommon, there is no grid current protection. It is
>>my opinion that a foolproof, frangible reasistor is a more reliable.
>>>
>
>Wouldn't it be more logical to ask *why* the electronic protection
>shorts during a major fault, and fix that problem?
Superb question, Mr. White. My guess is that there is too much current
during one type of fault for a 1 or so ampere rated transistor to handle
without developing a C-E short.
>
>Removing the electronic protection means there is NO protection against
>lesser faults,
Not if one adds a fusing element in its stead, which is what I said.
> such as incorrect loading and/or persistent overdriving.
On page 17 in the September 1990 issue of *QST*, there is a photograph of
an oxide-cathode / gold-plated grid amplifier tube which was overdriven
to the point of failure. The grid's gold-plating is in mint condition.
The cathode's strontium oxide and barium oxide coating has exfoliated,
causing a grid-cathode short. . If you arrive at page 17, page 16
has an example of the tendency of gold plating nearest the grounded end
of an 8877's grid to sputter/boil away, without visible damage to the
gold at the ungrounded end of the grid.
>These are equally capable of destroying the grid - it only takes a
>little longer.
>
>Rich's own web pages describe how to protect the grid and anode current
>meters from the current surges in "major faults", using shunt diodes.
>Since the fault protection circuit is essentially connected across the
>grid current metering resistor, the grid current sensing circuit ought
>to be as safe as the meters.
Grid current sensing circuits appear to work ok. The problem is with
shorting of the transistor that the sensing circuit controls -- the
transistor whose job it is to interrupt the flow of grid current during a
major glitch.
>
>If it doesn't work out that way, then there's a design error.
Agreed. In the Kenwood TL-922, a 1A-rated choke in the DC grid current
path to chassis-gnd. has been known to open during an event which
typically damages the amplifier's vhf suppressor resistors and arcs the
Tune-C, or the bandswitch. How much current does it take to melt the
Cu-wire in a 1-amp RFC? My unexpert guess is that if there's enough
grid current - from whatever - to burn out a 1A choke, there's probably
enough current to torpedo a 1A transistor.
>The meter protection diode clamps at abut 1 volt, so a surge limiting resistor
>between that diode and the base of the sensing transistor ought to do it.
agreed
>Another possibility is voltage drops in the chassis itself, if the
>emitter of the sensing transistor is connected to chassis at a different
>place than the diode.
... seems a bit unlikely with an aluminum chassis and a reasonable
layout.
>
>Find the error, fix it, and you can have routine fault protection that
>will also survive a major fault.
Surely, but it might take a transistor that can handle more than an
ampere or two. However, if one has a 20 or so Watt rated grid, and there
is a 1/2 Watt rated carbon film resistor in series with the DC grid
current path, which seems more likely to blow first? Sometimes, mo'
simpler is mo' betta.
Rich...
R. L. Measures, 805-386-3734, AG6K, www.vcnet.com/measures
--
FAQ on WWW: http://www.contesting.com/ampfaq.html
Submissions: amps@contesting.com
Administrative requests: amps-REQUEST@contesting.com
Problems: owner-amps@contesting.com
Search: http://www.contesting.com/km9p/search.htm
|